On Wed, Mar 6, 2013 at 09:16:59AM -0500, Tom Lane wrote:
> Simon Riggs <simon@2ndQuadrant.com> writes:
> > On 5 March 2013 22:02, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> >> FWIW, my opinion is that doing anything like this in the planner is
> >> going to be enormously expensive.
>
> > As we already said: no MVs => zero overhead => no problem.
>
> Well, in the first place that statement is false on its face: we'll
> still spend cycles looking for relevant MVs, or at least maintaining a
> complexly-indexed cache that helps us find out that there are none in
> a reasonable amount of time. In the second place, even if it were
> approximately true it wouldn't help the people who were using MVs.
>
> > It costs in
> > the cases where time savings are possible and not otherwise.
>
> And that is just complete nonsense: matching costs whether you find a
> match or not. Could we have a little less Pollyanna-ish optimism and
> a bit more realism about the likely cost of such a feature?
Should we add this to the TODO list as a possibility?
-- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB
http://enterprisedb.com
+ It's impossible for everything to be true. +